After Government Turbulence, GPT‑5.6 and ChatGPT Work Crash‑Land in the Office—and Workers Must Catch Up

• From regulatory drama to public rollout
• Tech companies embrace GPT‑5.6
• Inside ChatGPT Work’s agentic office
• The promise—and the paradox—for workers
• A crowded race for the AI office

After Government Turbulence, GPT‑5.6 and ChatGPT Work Crash‑Land in the Office—and Workers Must Catch Up
OpenAI’s latest AI push is colliding with everyday office life, as powerful new models and “agentic” tools arrive just weeks after regulators temporarily slowed their debut. Now businesses are racing to understand what work looks like when software can run projects for hours on its own.

§From regulatory drama to public rollout

GPT‑5.6 first appeared in a tightly controlled “limited preview,” accessible only to government‑approved organizations amid concerns over its cybersecurity capabilities and potential misuse. After the Trump administration granted a green light, OpenAI began a global public rollout of GPT‑5.6 and its three‑tiered family of models—Sol, Terra, and Luna.

On the same day, OpenAI introduced ChatGPT Work, an AI agent that “can gather context from the apps, files, and workflows you choose and create finished materials such as documents, spreadsheets, presentations, and web apps.” The launch was billed internally as a “massive day” that brought “agents for everyone in the new ChatGPT app Work and Codex modes.”

§Tech companies embrace GPT‑5.6

OpenAI quickly positioned GPT‑5.6 as its flagship, calling it its “strongest cybersecurity model yet” and highlighting major gains in efficiency and coding performance. Microsoft made GPT‑5.6 the “preferred model in Microsoft 365 Copilot,” so users of Word, Excel, PowerPoint, Chat and Cowork can “create higher‑quality work products with less effort.” CEO Satya Nadella said the update brings “stronger reasoning and higher‑quality outputs without sacrificing efficiency” across Microsoft’s ecosystem.

Cloud platforms and tools followed. Amazon Bedrock announced GPT‑5.6 Sol, Terra and Luna as “generally available,” offering “three tiers of intelligence, from flagship reasoning to fast inference.” Perplexity’s team said the family “sits on the Pareto frontier of our agentic research evals: higher accuracy at lower cost.”

OpenAI’s own apps are being re‑engineered around the models. Codex is being merged into a “super app for work” built on ChatGPT, as the company seeks to become “an indispensable part of life for millions of workers.” GPT‑5.6 is also now “the preferred model in Microsoft 365 Copilot” and powers ChatGPT Work’s long‑running, multi‑step automation.

§Inside ChatGPT Work’s agentic office

ChatGPT Work is pitched as an agent that “can take action across your apps and files, stay with a project for hours if needed, and turn a goal into finished work.” It can automate workflows—such as transforming customer research into a campaign brief and then localized marketing assets—while waiting for humans to “approve important actions.”

OpenAI is also rolling out ChatGPT Sites, which can turn “a prompt, file, or rough idea into a dashboard, project tracker, report, prototype, or lightweight app” and publish it with a URL. The company says its updated desktop app unifies chat, Codex and Work, so teams can delegate tasks, spin up separate chats to brainstorm, and let agents operate autonomously in a browser or across local files.

Executives at OpenAI argue this marks a shift in knowledge work. One internal framing shared publicly describes jobs moving from “handling individual tasks” to “tending systems with AI loops” that sweep email, decide what deserves attention, and research context before surfacing answers. Greg Brockman says ChatGPT Work “brings agents to consumer scale” and makes it “incredibly joyful to just ask any question about the business and have it be thoroughly researched and answered.”

§The promise—and the paradox—for workers

For many, the potential upside is efficiency. OpenAI claims GPT‑5.6 is “orders of magnitude more efficient and cost‑effective” than prior models, with Sol described by Sam Altman as half the price and roughly twice as token‑efficient as a key rival for similar tasks. Early health benchmarks even suggest “physicians found fewer flaws in GPT‑5.6 responses than physician‑written responses,” prompting some to call the update a “major step forward for health.”

But the new power also brings complexity. Axios notes that users now face “a jumble of new model names, reasoning levels and subscription tiers” and must figure out which combination of Sol, Terra or Luna, plus reasoning settings, fits their work and budget. The piece dubs it a “paradox of choice,” where non‑coders can get “more work better, faster and cheaper — if those workers can determine the ‘just right’ settings for their use case.”

Office managers must also decide how much autonomy to grant these systems. Ars Technica highlights a lingering concern in its piece titled “OpenAI wants its new tool to do your work for you and with you,” asking bluntly: “Give ChatGPT access to all your work tools; what could go wrong?”

§A crowded race for the AI office

The GPT‑5.6 family and ChatGPT Work arrive as competitors like Anthropic, Google, Meta, SpaceXAI and open‑source agents vie to become the default AI coworker. OpenAI is pitching Sol as its “workhorse” that outshines Anthropic’s models on coding and design benchmarks, with third‑party leaderboards now placing GPT‑5.6 Sol first overall on frontend design with an Elo of 1353.

Usage numbers suggest the bet is working, at least initially. OpenAI says it has seen a 2.5x surge in usage of its agentic products in a single week and 8 million active users across Codex and ChatGPT Work, with limits repeatedly reset to keep up with demand. As one Business Insider headline puts it, the company is “making its biggest play for the office” by folding Codex into ChatGPT and banking that workers will adapt quickly enough to let GPT‑5.6 do their work “for you and with you.”

Continue reading https://foxvector.com/stories/019f6236-676e-1751-72ef-1bec67003e58